data-api-builder
data-api-builder copied to clipboard
Use Memory Stream for Serialization
Why make this change?
We have recently seen an out of memory error coming from the line of code that serializes the column data.
What is this change?
Here, we are using a memory stream in the serialization process to avoid any out of memory errors.
There is a small performance overhead in using a memory stream - and I would like your opinion on whether that could be significant or not. The overhead could be negligible compared to allocating a large string, however, if you think performance would be an issue, we could change the code to only use the memory stream if the data is larger than a threshold
How was this tested?
- [ ] Integration Tests
- [ ] Unit Tests
Sample Request(s)
- Simple query request
- Stored procedure call
- Embedded stored procedure calls
Thank you for this contribution. I believe this is a solid upgrade.